Electoral Re-registration, Disenfranchisement and Public Service Provision

Advisor: Claudio Ferraz

This paper examines whether reforms aimed at reducing electoral fraud can have the unintended effect of disenfranchising poorer citizens and, as a consequence, affect public services delivered to poorer households. We exploit a large program of re-registration of voters in Brazil’s municipalities where the electoral commission suspected the presence of electoral fraud. Using a difference-in-difference strategy, we compare the electoral registration and turnout in 1186 Brazillian municipalities that went through electoral revision with those that did not in elections prior to the re-registration and after the change. We find that the program reduced registration rates by 10 percentage points and participation rates by 5 percentage points, specially in municipalities with low education levels and with low media penetration. Moreover, we find that the newly elected mayors responded to this change by reducing public expenditure in areas that disproportionately benefits poor and uneducated voters (education and health). Finally, we show that the reduction in expenditures deteriorated the infrastructure of public schools and worsened health outcomes of less educated citizens.